2019 | 72 / Special Issue | 71–80 | 5 Figs. | 1 Tab. | 2 App. Tabs. | www.geologia-croatica.hr Journal of the Croatian Geological Survey and the Croatian Geological Society Article history: Manuscript received April 11, 2019 Revised manuscript accepted July 08, 2019 Available online December 20, 2019 Keywords: rock slopes, hydropower engineering region, slope stability probability classification system, modification, Hoek-Brown strength criterion, CSMR system 1. INTRODUCTION Slope rock mass is a type of very complex material with time- space variability. Under the natural state, the rock mass not only has a long and complicated deformation history, but also includes many crisscrossed discontinuous planes such as joints and frac- tures after experiencing many orogenetic and tectonic move- ments. It has also usually been affected over a long period by many kinds of natural factors such as weathering and rainfall, as well as construction and other man-made factors. Under complex geological conditions, it is difficult to accurately determine the spatial and temporal distribution of rock mass properties except through careful investigation and testing. Therefore, it is difficult for any kind of mechanical model to describe its mechanical be- haviour in an all-round and accurate way. Pure theoretical calcu- lation and experimental analysis often fail to solve practical prob- lems. The problems often need geological engineers to make decisions on the base of their experience. Because slope rock and soil mass is extremely complex, we are still far from a full and perfect understanding of its geologi- cal characteristics, deformation, strength and mechanical prope- rties (CHEN, 2005). Therefore, the study of rock slopes is still in the process of continuous exploration and improvement based on experience. The stability assessment of rock slopes in a hydro- power engineering region is especially a very important and com- plicated issue. At present, rock mass classification systems pro- vide a good approach and have been widely applied in the stability assessment of rock slopes by many researchers because they can consider many geological factors that affect slope stabi­ Modification of slope stability probability classification and its application to rock slopes in hydropower engineering regions Li Xiu-Zhena,b*, Tan Rong-Zhia,b* and Gao Yanc a Chinese Academy of Sciences, Key Laboratory of Mountain Hazards and Earth Surface Process, 610041, Chengdu, China; (lxzljt@imde.ac.cn) b Chinese Academy of Sciences, Institute of Mountain Hazards and Environment, 610041, Chengdu, China; (corresponding author: tanrz@imde.ac.cn / lxzljt@imde.ac.cn) c Guizhou University, College of Resources and Environment Engineering, 550025, Guiyang, China doi: 10.4154/gc.2019.20 Abstract Stability assessment of rock slopes in hydropower engineering regions is an important and com- plex issue. Rock mass classification systems are a good approach because they can thorough- ly consider many factors influencing rock slope stability. The slope stability probability classifi- cation (SSPC) system is a novel method. However, it has two limitations when applied to rock slopes: 1) it is only suitable for slopes less than 45 m in height, and 2) there is great subjectivity and randomness in the estimation of intact rock strength. Therefore, this study presents two modifications of the SSPC system by adopting the Hoek-Brown strength criterion and an em- pirical formula for maximum slope height. Evaluation of results from of 34 typical rock slopes of the major hydropower engineering regions in China indicated that the accuracy rate of the mod- ified SSPC for stability evaluation of these slopes was 61.8%, and the accuracy for stability eval- uation of 10 slopes with non-structural control failure was 80%. The stability values of stable and unstable slopes obtained using the modified SSPC were different to those obtained using the Chinese Slope Mass Rating (CSMR) and modified CSMR systems. In addition, the identification accuracy rate of the modified SSPC was significantly higher than that of the CSMR and modi- fied CSMR. Therefore, the modified SSPC can be applied to hydropower engineering regions, providing a new means of rapidly evaluating the slope stability of high rock slopes (slopes > 45 m in height) in these regions. lity and obtain a quantitative empirical formula. Since the 1870s, many scholars have put forward various rock mass classification systems for rock slope stability evaluation (PANTELIDIS, 2009; RUSSELL et al., 2009; XIAO, 2007; ZHENG et al., 2016), such as Rock Mass Rating (RMR) by BIENIAWSKI (1974), Slope Mass Rating (SMR) by ROMAN (1985), Rock Mass Strength (RMS) by SELBY (1980), Slope Rock Mass Rating (SRMR) by ROBERTSON (1988), Geological Strength Index (GSI) by HOEK et al. (1988, 2002) and CSMR for slopes in hy- dropower engineering region by CHEN et al. (1997). SHI et al. (2005) proposed the Highway Slope Mass Rating (HSMR) sys- tem for rock slopes of mountain highways based on the SMR. WU et al. (2005) proposed a General Slope Mass Rating (GSMR) system applicable to the evaluation of rock slope stability based on a large number of practical engineering research projects. LI et al. (2010) proposed a modified CSMR using a continuous func- tion to modify the systematically modify quantitative parameters in CSMR. DAFTARIBESHELI et al (2011) applied fuzzy set theory to the RMR system and presented a Fuzzy Slope Mass Rating (FSMR) system. All of these classification systems pro- vide an important means for the rapid evaluation of rock slope stabi lity (FRANCIONI et al., 2018; MORALES et al., 2019). However, most of the aforementioned systems for rock slope stability classification are based on a single weight value to evalu­ ate slope stability, and the failure mechanisms and modes of rock slopes are not strictly considered. For example, the slope stability of structural control failure is mainly affected by the structural plane condition and the relationship between the structural plane G eo lo gi a C ro at ic a Geologia Croatica 72 / Special Issue72 and slope orientation. However, the slope stability of non-struc- tural control failure is mainly affected by the shear strength of the slope rock mass and height. In addition, the existing classifi- cation systems do not clearly distinguish an exposure rock mass and a slope rock mass, the characteristics of which may be quite different due to the influences of weathering and excavation (HACK, 2002). Hack put forward the Slope Stability Probability Classifica- tion (SSPC) system in 1998 based on the aforementioned issues and the shortcomings of the existing slope stability classification systems (HACK, 2002; HACK et al., 2003). The SSPC system resulted in great progress in the evaluation of the stability of rock slopes. For example, the adoption of a continuous formula during the calculation process ensures non­step classification results. The stability evaluation of rock slopes has been divided into ori- entation-dependent stability and orientation-independent stabi lity according to slope failure types. The evaluation result depends on the probability of slope failure in different modes, but not on a single weight value. The SSPC system has been applied and de- veloped in the study of highway slopes in Spain for four years, and has also been applied in Austria, South Africa, New Zealand, China, and the Netherlands and has achieved good results (DAS et al., 2010; HACK et al., 2003; LI and XU, 2016; LINDSAY et al., 2000; LINDSAY et al., 2001; CANAL et al., 2016). The empirical formula of the SSPC system is mainly based on the statistical analysis of 184 highway slopes with a slope height less than or equal to 45 m (HACK et al., 2003), therefore, this system may be more suitable to the stability evaluation of rock slopes with a slope height less than 45 m. In addition, SSPC emphasizes the influence of weathering and excavation on slope stability and pays relatively little attention to the intact rock strength compared to previous slope stability classification sys- tems. In the SSPC system, the parameter of intact rock strength is mainly estimated by field observation and a simple hammer test, which increases its subjectivity and randomness. LINDSAY et al. (2000) also noted that this estimating method of intact rock strength is the major shortcoming of the SSPC system. The SSPC system cannot be directly used to evaluate the sta- bility of rock slopes in a hydropower engineering region as slope heights are generally greater than 45 m. A modified method of shear strength and maximum slope height of a rock slope in the SSPC system was proposed in this study adopting the Hoek- Brown strength criterion and an empirical formula of maximum slope height, based on the limitations of slope stability evaluation in the SSPC. An analysis of some case studies showed that the modified SSPC can be used for probability evaluation of rock slope stability in a hydropower engineering region and can pro- vide a new means for rapid evaluation of rock slope stability. 2. THE SSPC CLASSIFICATION 2.1. Overview The method considers three kinds of rock mass including expo- sure rock mass (ERM), reference rock mass (RRM) and the slope rock mass (SRM), obtains rock mass parameters based on inves- tigation and testing on the slopes in the field, and identifies the possible failure mode and instability probability according to fail- ure modes and mechanisms of the rock slopes. The ERM is the rock mass in the exposure; the RRM is the rock mass in an ima- ginary, unweathered, and undisturbed condition prior to excava- tion; and the SRM is the rock mass in which the existing or new slope is to be situated. Compared to the SMR classification systems, the SSPC method has made great progress in the stability identification of rock slopes. The main advantages of the method include: (1) it has strong operability, and its evaluation parameters are easy to obtain in the field; (2) the continuous formulae are adopted in the calculation process, which guarantees the non-step property of the graded results; (3) evaluating orientation-independent slope stability is based on the classical slope stability analysis method, evaluating orientation-dependent slope stability embodies the controlling effect of structural surface condition and features val- ues on the slope stability; (4) evaluation result depend on the probability valu es that the slope may occur in different failure modes, and does not only depend on a rating weight value such as the SMR classification systems. 2.2. Basic theory The concept of the SSPC system is based on the following three aspects (HACK, 2002). (1) A three­step classification system is introduced to de- scribe the exposure rock mass, the reference rock mass, and the slope rock mass (Fig. 1). (2) The slope stability is determined by the probable occur- rence of different failure mechanisms instead of a single weight value. (3) Unambiguous and simple procedures for data collection in the field. The assessment procedure of the method can be seen in Figure 2. 2.3. Evaluation indexes The evaluation indexes used in the SSPC system mainly include intact rock strength, orientation, spacing and the number of dis- continuity sets, shear strength characteristics of the discontinui- ties. The acquisition and quantification of intact rock strength in this classification system are mainly estimated by field observa- tion and a simple hammer test. The relationship between the ori- entation of discontinuity and the orientation of slope determines the failure mechanism and failure mode of the rock slope. In the SSPC system, the influence of the discontinuity orientation on the slope stability is mainly reflected in the change in the appar- ent dip (AP) of the structural surface. AP can be calculated using the following formula: arctan(cos( ) tan )s j jAP α α β= − ⋅ (1) Figure 1 Sketch of exposures in rock masses of various degrees of weathering and different types of excavation indicating the concept of the ‘reference rock mass’ (HACK, 2003). G eologia C roatica Xiu-Zhen et al.: Modification of slope stability probability classification and its application to rock slopes in hydropower engineering regions 73 In this formula, αs is the slope direction, αj is the discontinu- ity dip direction, and βj is the discontinuity dip angle. In the SSPC system, the combination of the spacing and the number of discontinuities is mainly quantified by three groups of discontinuities with the smallest spacing, according to the dia- grammatic method proposed by TAYLOR (1980). The conditions of the discontinuities determine their shear strength. The charac- teristics of the discontinuities are determined by four main fac- tors: large-scale roughness (Rl), small-scale roughness (Rs), infill material (Im), and karst (Ka). A discontinuity condition factor (TC) can be determined by a multiplication of the four factors as follows: * *Im*TC Rl Rs Ka= (2) 2.4. Evaluation rules The slope stability of the SSPC system is determined using two analyses according to the failure mechanism and main control factors of the rock slopes: one is related to the orientation of the discontinuities and the slope (orientation-dependent stability), and the other is unrelated to the orientation of the discontinuities and the slope (orientation-independent stability). The former is for stability analysis of rock slopes of structural control failure, while the latter is for stability analysis of rock slopes of non-struc- tural control failure. 1. Orientation-dependent stability assessment This type of slope stability analysis mainly considers the condi- tion of discontinuity planes, the relationship between dip direc- tion and angle of discontinuity planes and dip direction and angle of slopes. According to the failure criteria of sliding and dump- ing, the failure probability of the rock slope in different modes is analysed, and the maximum probability is determined as the pos- sible failure probability and the corresponding failure mode is taken as the possible failure mode of the rock slope. For sliding failure, the SSPC system built a graph between the condition pa- rameters of discontinuous plane and the apparent dip angle of discontinuous plane as a criterion to evaluate the stability proba- bility of the slopes. For toppling failure, the relationship between the condition parameters of discontinuous plane and the apparent dip angle of discontinuous plane and slope angle is established as a criterion to evaluate the stability probability of slopes. 2. Orientation-independent stability assessment This type of slope stability analysis adopts a linear shear plane model which follows the Mohr-Coulomb failure criterion. Firstly, by determining the cohesion and internal friction angle of the slope rock mass, the maximum stability slope height is calculated. Secondly, the ratio of the maximum stable slope height to the ac- tual slope height and the ratio of the internal friction angle of the rock mass to the actual slope angle are calculated. Finally, ac- cording to the linear shear plane failure model, the possible fail- ure probability of rock slopes can be obtained by means of the related figures published in HACK et al. (2003). Detailed descriptions and related figures regarding the SSPC method are available in HACK (2002) and HACK et al. (2003). 3. MODIFICATION OF THE SSPC FOR ROCK SLOPES IN HYDROPOWER ENGINEERING REGIONS 3.1. Limitations of the SSPC As previously mentioned, the empirical formula in the SSPC sys- tem (such as the calculation formula of shear strength and the maximum slope height of a rock mass) is mainly based on the analysis of highway slopes in Spain; thus, it is more suitable in the stability evaluation of slopes below 45 m in height. In addi- tion, compared to previous slope stability probability classifica- Figure 2. Flow diagram of the three-step concept of the SSPC system (HACK, 2003). G eo lo gi a C ro at ic a Geologia Croatica 72 / Special Issue74 tion systems, the SSPC system emphasizes the influence of weathering and excavation on slope stability, while the intact rock strength is estimated via field observation and a simple hammer test. The estimation of the strength is strongly subjective in the SSPC system (LINDSAY et al., 2001). In 2016, application of the SSPC method in the stability as- sessment of highway slopes in China obtained good results (LI & XU, 2016). The original plan was to use the SSPC method to assess the stability of hydropower engineering slopes. However, it was discovered that the SSPC system isn’t very suitable for the slopes in hydropower regions, due to the greater height of these slopes (generally more than 45 m). Based on the limitations of slope stability evaluation in the SSPC system, a modifica- tion method of shear strength and maximum slope height of rock slopes of non-structural control failure was proposed adopting the Hoek-Brown strength criterion and an empirical formula of maximum slope height, while the SSPC system was still used to evaluate the stability of the rock slope of structural control fail- ure. The specific modification methods are described below. 3.2. Modification of SRM strength in the SSPC The modification of the shear strength of the SRM of non­struc- tural control failure is mainly based on the relatively perfect Hoek-Brown empirical strength criterion (HOEK & BROWN, 1980, 1988; HOEK, 1990; HOEK et al., 2002). The calculation formula of parameters c′ and φ′ of the equivalent Mohr­Coulomb rock mass strength in a different range of slope height stress can be derived from the linear Mohr-Coulomb failure criterion and related rock mass parameters, including the geological strength index (GSI), lithological coefficient (mi), and the uniaxial com- pressive strength (σci) (HOEK et al., 2002) as follows: ' ' 3 ' 1 3 ' 1 3 [(1 2 ) (1 ) ] (1 )(2 ) ( ) 1 (6 ( ) ) / ((1 )(2 )) ci b n a b n a b b n a s a m c a a s m am s m a a σ σ σ σ − − + + − = × + + + + + + + (3) ' 1 ' 1 3 ' 1 3 6 ( ) sin [ ] 2(1 )(2 ) 6 ( ) a b b n a b b n am s m a a am s m σ ϕ σ − − − + = + + + + (4) where ' 3 3max /n ciσ σ σ= , ciσ is the uniaxial compressive strength of the rock. ' 3maxσ is the upper limit of the stress range calculated using the Bishop method under a different slope height, and it can be obtained using the following formulas: 0.91' ' 3max ' 0.72 cm cm H σ σ γσ −   =      (5) 1 ' ( 4 ( 8 ))( / 4 ) 2(1 )(2 ) a b b b cm ci m s a m s m s a a σ σ −+ − − + = ⋅ + + (6) where γ is the bulk density of the rock mass, mb is the material parameter of the rock mass, and a and s are parameters of the rock mass that can be obtained using the following formulas: 100exp 28 14b i GSIm m D − = ⋅  −  (7) ( )/15 20/31 1 2 6 GSIa e e− −= + − (8) 100exp 9 3 GSIs D − =  −  (9) The three indexes of the rock mass; disturbance coefficient D, geological strength index GSI and lithological coefficient mi in formulae (7)–(9) can be determined based on the correspond- ing charts in MARINOS & HOEK (2000) and CHEN et al. (2005) which provide further detail. 3.3 Modification of slope height in SSPC For a slope of non-structural control failure, HUANG (1994) cal- culated the critical slope of a homogeneous limited rock slope with different slope heights when the safety factor was 1, using the equilibrium limit analysis method. The empirical formula of maximum slope height was obtained based on the known litho- logy, the rock mass structure, rock strength, and rock weight us- ing a regression and nonlinear method according to the analysis results (HUANG, 1994) as follows: i( 0.0003 m 0.0483)1.5 max (0.00651 0.00037 ) ( ) GSIci iH m e σ γ − × += + × × × (10) where maxH is the critical height (m) representing the slope height when the tangent value of the slope angle is closer to in- finity (the slope angle is near 90°), and the safety factor is equal to 1. The result of formula (10) better represents the real condi- tions of the slope (HUANG, 1994) and has been verified by en- gineering examples. Therefore, this formula was used to calculate the maximum slope height of the slopes in this study. 4. PRELIMINARY APPLICATION OF THE MODIFIED SSPC TO A ROCK SLOPE IN A HYDROPOWER ENGINEERING REGION 4.1. Data source Since the 1980s, numerous high and steep slope problems have occurred in China with the construction of many important hy- dropower projects. CHEN (2004) took part in many scientific re- search projects and advisory work regarding high and steep slope problems of the projects at a national and ministerial level. CHEN and his team (2004) established a database including 115 slopes in hydropower engineering regions during his implementation of the research projects. Most of the slopes in the database have been subject to special investigation and research studies, and there are clear conclusions regarding their geometric characteristics, engi- Table 1. Comparison of the accuracy of the stability identification systems applied to the 34 rock slopes. Correct number and accuracy rate of evaluation CSMR system Modified CSMR system Modified SSPC system 34 slopes Correct number 14 15 21 Accuracy 41.18% 44.12% 61.76% 10 slopes of non-structural control failure Correct number 5 7 8 Accuracy 50% 70% 80% G eologia C roatica Xiu-Zhen et al.: Modification of slope stability probability classification and its application to rock slopes in hydropower engineering regions 75 neering geological characteristics, slope structure, discontinuity conditions, and stability conditions. In particular, for some of the rock slopes, there are complete and detailed stability classifica- tion indexes, such as uniaxial compressive strength, excavation methods, weathering strength, etc. Therefore, 34 slopes in hydro- power engineering regions with detailed evaluation indexes in the database were used to create case studies here (Appendix Ta- bles, Table A1). 4.2. Process and steps The detailed analysis and calculation steps of the modified SSPC system are as follows: (1) First, the lithology coefficient mi is determined according to the type of rock slope; the value of geological strength index GSI is comprehensively determined according to the rock type, weathering degree, rock mass structure and the conditions of dis- continuities; the value of D is determined by interpolation in the range of 0 to 1 according to the slope excavation method; and the weight γ of different rocks is determined by referring to the rele­ vant manual of rock mechanics and the results of laboratory tests in Chen’s database previously mentioned. (2) Second, the value of the cohesive force c′ and internal friction angle φ′ of the rock mass of different slopes is calculated using the free Roclab software (http://roclab.software.informer. com/), according to the Hoek–Brown strength criterion, and based on the known intact rock strength σc, geological strength index GSI, lithology coefficient mi, and disturbance coefficient D. Then the ratio of internal friction angle and actual slope φ′/βs is calculated (Appendix Tables, Table A2). (3) Third, the maximum slope height Hmax of different slopes can be obtained according to formula (10), and the ratio of the maximum slope height of the stable slope to the real slope height is calculated (Appendix Tables, Table A2). (4) Finally, the stability probability of the rock slope is ob- tained according to the value of φ′/βs and Hmax/H, and referring to the original SSPC system; then, the slope stability is evaluated according to the following criteria: When the slope stability probability SP ≤ 40%, the slope is unstable; when 40% < SP ≤ 70%, the slope is partially unstable; and when SP ≥ 70%, the slope is stable. 4.3. Results Characteristic information of 34 rock slopes was extracted from the slope engineering database of China’s key hydropower engi- neering region, which was created by CHEN (2004), such as li- thology, slope structure, conditions of discontinuities, excavation method, and slope types (Appendix Tables, Table A1). The stabi- li ty of the rock slopes was evaluated using the aforementioned modified SSPC system, and the evaluation results are shown in Table A2 of Appendix Tables. Table A2 and Fig. 3 show that the evaluation accuracy rate of the 34 slopes using the modified SSPC system is 61.76% (21 are correct and 13 are incorrect). The slopes with the correspond- ing stability level have the largest proportion in each stability class (Fig. 4). The evaluation accuracy rate of the 10 slopes of non­structural control failure reaches 80%. Only the evaluations of the No. 2 and No. 17 slopes are incorrect, and the evaluation results of the other 8 slopes are consistent with the actual stabil- ity (Table A2). 5. CONCLUSION AND DISCUSSION The SSPC system was a slope stability probability classification system proposed by HACK in 1998. Via a three-step analysis method, it considered three types of rock mass, ERM, RRM, and SRM, and analysed the failure probability in different failure modes via field investigation, calculating various parameters of rock mass, and combined with the failure mode and failure mecha- nism of the rock slope, evaluated the potential failure mode and failure probability. The SSPC system has resulted in great pro- gress in the evaluation of rock slope stability compared to other classification systems. However, there are two limitations of this system: 1) it is more suitable for stability evaluation of slopes less than 45 m in height, and 2) there is a subjectivity in its compres- sion strength estimation of intact rock. Therefore, the SSPC sys- tem may not be very suitable to the evaluation of rock slopes in hydropower engineering regions considering that most of the slope heights in such areas exceed 45 m. Based on this, a modified method of shear strength and maxi­ mum slope height of a rock slope in the SSPC system was pro- posed here adopting the Hoek-Brown strength criterion and an empirical formula of maximum slope height. The stability of 34 typical rock slopes in hydropower engineering regions in China was evaluated using the modified SSPC system. The evaluation results indicated that the accuracy of the modified SSPC system Figure 3. Distribution of 34 slopes on the probability map of orientation-inde- pendent stability using the modified SSPC system. Figure 4. Percentage of slopes of different stability classes in each stability class G eo lo gi a C ro at ic a Geologia Croatica 72 / Special Issue76 for stability evaluation of these slopes was 61.76% and the accu- racy for stability evaluation of 10 slopes of non-structural control failure was 80% (Table 1). To further compare the application effectiveness of the modi­ fied SSPC system, the stability of the aforementioned 34 rock slopes was completed grading the evaluation based on the CSMR system put forward by CHEN et al. (1997) and the modified CSMR system put forward by LI et al. (2010). More details on these two methods are available in CHEN et al. (1997) and LI et al. (2010). The evaluation results are shown in Fig. 5 and Table 1. Table 1. shows that the value differences of slope stability eva­ luation obtained using the CSMR and modified CSMR systems are not significant, while the probability values of slope stability ob- tained using the modified SSPC system are significantly different. Slopes with different degrees of slope stability degree can be better separated (HACK et al., 2002). Moreover, the identification accu- racy rate of the modified SSPC system is obviously higher than that of the CSMR and modified CSMR systems (Table 1). Therefore, the modified SSPC system can be applied to sta- bility probability classification of rock slopes in a hydropower engineering region. It can provide a new effective means for the rapid stability evaluation of rock slopes in a hydropower engi- neering region with heights exceeding 45 m. However, it should be noted that the stability evaluation of structural control slopes in this study cannot further calculate and validate analysis since the field investigation data of the 34 slopes in the database are neither very detailed nor complete, particu- larly the occurrence, number and spacing of discontinuities. In practice, according to the SSPC system, the analysis of orienta- tion-dependent and orientation-independent stability should both be conducted and the lesser probability be taken as the final as- sessment result for a rock slope. ACKNOWLEDGMENT This research was supported by the National Key Basic Research Program of China (No. 2015CB452704), Natural Science Foun- dation of China (No.Y8K1200200), and the open foundation of the State Key Laboratory of Geohazard Prevention and Geoen- vironment Protection (SKLGP2013K023). In particular, I would like to thank academician CHEN ZUYU and Professor WANG YUJIE of the China Water Resources & Hydropower Science Research Institute for providing the slope database for the water conservancy and hydropower projects used in this study. REFERENCES BIENIAWSKI, Z.T. (1973): Engineering classification of jointed rock masses.– The Ci vil Engineer in South Africa, 15/12, 343–353. CANAL, A. & AKIN, M. (2016): Assessment of rock slope stability by probabilistic-based slope stability probability classification method along highway cut slopes in Adil- cevaz-Bitlis (Turkey).– Journal of Mountain Science, 13/11, 1893–1909. doi: 10.1007/s11629-016-3954-y CHEN, Z.Y., WANG, X.G., YANG. J., JIA, Z.X. & WANG, Y.J. (2005): Rock slope sta- bility analysis: principle, method, and procedure.– Beijing: China Water Power Press (in Chinese). DAFTARIBESHELI, A., ATAEI, M. & SERESHKI, F. (2011): Assessment of rock slope stability using the Fuzzy Slope Mass Rating (FSMR) system.– Applied Soft Com- puting, 11/8, 4465–4473. doi: 10.1016/j.asoc.2011.08.032 DAS, I., SAHOO, S., WESTEN, C.V., STEIN, A. & HACK, R. (2010): Landslide suscep- tibility assessment using logistic regression and its comparison with a rock mass classification system, along a road section in the Northern Himalayas (India).– Geo­ morphology, 114/4, 627–637. doi: 10.1016/j.geomorph.2009.09.023 FRANCIONI, M., STEAD, D., SCIARRA, N. & CALAMITA, F. (2018): A new approach for defining Slope Mass Rating in heterogeneous sedimentary rocks using a com- bined remote sensing GIS approach.– Bulletin of Engineering Geology and the En- vironment, 1–22. doi: 10.1007/s10064-018-1396-1 HACK, R. (2002): An evaluation of slope stability classification.– In: Proceedings of the Eurock, Portugal, 3–22. HACK, R. & PRICE, D. (1993): A rock mass classification system for the design and safety analysis of slopes.– Eurock, 803–810. HACK, R., PRICE, D. & RENGERS, N. (2003): A new approach to rock slope stability – a probability classification (SSPC).– Bulletin of Engineering Geology and the En- vironment, 62, 167–184. HOEK, E. (1990): Estimating Mohr-Coulomb friction and cohesion values from the Hoek– Brown failure criterion.– International Journal of Rock Mechanics & Mining Sci- ences & Geomechanics, 27/3, 227–229. HOEK, E. & BROWN, E.T. (1980): Empirical strength criterion for rock masses.– Journal of Geotechnical and Geoenvironmental Engineering, 106/15715, 1013–1035. HOEK, E. & BROWN, E.T. (1988): The Hoek–Brown failure criterion – a 1988 update.– Journal of Heuristics, 16/2, 167–188. HOEK, E., CARRANZA-TORRES, C. & CORKUM, B. (2002): Hoek–Brown failure criterion-2002 edition.– In: Proceedings of the Fifth North American Rock Me- chanics Symposium, Toronto, 1–6. (c) modified SSPC system Figure 5. Comparison of the different stability identification systems (a–c) ap- plied to the 34 rock slopes. (b) modified CSMR system (a) CSMR system G eologia C roatica Xiu-Zhen et al.: Modification of slope stability probability classification and its application to rock slopes in hydropower engineering regions 77 HUANG, G.X. (1994) A proposed susceptibility­index and influence zone of rock ava- lanche.– Dissertation, National Central University (in Chinese). LI, X.Z., KONG, J.M. & WANG, C. (2010): Modification of rock slope stability clas- sification systems by continuous functions and its application.– Chinese Journal of Rock Mechanics and Engineering, 29/Supp.1, 3439–3446 (in Chinese). LI, X.Z. & XU, Q. (2016): Application of the SSPC method in the stability assessment of highway rock slopes in the Yunnan province of China.– Bulletin of Engineer- ing Geology and the Environment, 75/2, 551–562. LINDSAY, P., ANDERSON, J., BOURKE, F., CAMPBELL, R.N. & CLARKE, L. (2000): Predicting slope stability in open pit gold and coal mines.– In: New Zea- land Minerals and Mining Conference Proceedings, New Zealand, 29–31. LINDSAY, P., CAMPBELL, R.N., FERGUSSON, D.A., GILLARD, G.R. & MOORE, T.A. (2001): Slope stability probability classification, Waikato Coal Measures, New Zealand.– International Journal of Coal Geology, 45/2, 127–145. doi: 10.1016/S0166-5162(00)00028-8 LYSANDROS, P. (2009): Rock slope stability assessment through rock mass classification systems.– International Journal of Rock Mechanics and Mining Sci- ences, 46, 315–325. MARINOS, P. & HOEK, E. (2000): A geologically friendly tool for rock mass strength estimation. In: Proceeding on Geotechnical and Geological Engineering (Geo- Eng2000, Melbourne, Australian). MORALES, M., PANTHI, K.K. & BOTSIALAS, K. (2019): Slope stability assessment of an open pit mine using three-dimensional rock mass modelling.– Bulletin of Engineering Geology and the Environment, 78/2, 1249–1264. ROBERTSON, A.M. (1988): Estimating weak rock strength.– In: Proceedings of the SME Annual Meeting, Phoenix, Arizona, 1–5. ROCLAB software (http://roclab.software.informer.com/) ROMAN, M. (1985): New adjustment ratings for application of Bieniawski classifica- tion to slopes.– In: Proceedings of the International Symposium on Role of Rock Mechanics, Zacatecas, Mexico, 49–53. RUSSELL, C.P., SANTI, P.M. & HUMPHREY, J.D. (2009): Modification and statisti- cal analysis of the Colorado rockfall hazard rating system.– Engineering Geology, 104/1, 55–65. SELBY, M.J. (1980): A rock mass strength classification for geomorphic purposes: with tests from Antarctica and New Zealand.– Zeitschrift für Geomorphologie, 24, 31–51. SHI, Y.C., WANG, Z.W., WAN, G.R., WANG, Z.Y., CHEN, Q.Y. & TANG, S.C. (2005): Study of mountain highway slope mass rating.– Chinese Journal of Rock Mecha- nics Engineering, 24/6, 939–944. TAYLOR, H.W. (1980): A geomechanics classification applied to mining problems in the Shabanie and King mines, Zimbabwe.– Dissertation, University of Rhodesia. WU, D.B. & XU, W.Y. (2005): GSMR method for determining rock slope mechanical parameters.– Rock and Soil Mechanics, 26/9, 1421–1426. XIAO, G.F. (2007): Study on classification method of rock slope stability in highway construction in mountainous region.– Dissertation, Institute of Rock and Soil Me- chanical, the Chinese Academy of Sciences (in Chinese). ZHENG, J., ZHAO, Y., LÜ, Q., DENG, J., PAN, X. & LI, Y. (2016): A discussion on the adjustment parameters of the slope mass rating (SMR) system for rock slopes. – Engineering Geology, 206, 42–49. doi: 10.1016/j.enggeo.2016.03.007 G eo lo gi a C ro at ic a Geologia Croatica 72 / Special Issue78 Ta bl e A1 . B as ic c ha ra ct er is tic s a nd d at a of th e 34 sl op es in h yd ro po w er e ng in ee rin g re gi on . N um be r Sl op e na m e Li th ol og y Ro ck m as s st ru ct ur e di sc on tin ui ty co nd iti on Ex ca va tio n m et ho d Sl op e ty pe σ c (M Pa ) G SI m i D Γ （ M N /m 3 ） 1 Ta ng Y an gu an g la nd sl id e at th e Zh ex i H yd ro po w er S ta tio n st ro ng ly w ea th er ed fi ne sa nd st on e in fil l s an dy sl at e bl oc ky po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 35 45 7 0. 00 0. 02 8 2 1# la nd sl id e at th e Li jia xi a D am st ro ng ly w ea th er ed c ho ris m ite a nd sc hi st lo os e po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 45 20 29 0. 00 0. 02 7 3 2# la nd sl id e at th e Li jia xi a D am st ro ng ly w ea th er ed c ho ris m ite a nd sc hi st lo os e po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 48 20 29 0. 00 0. 02 7 4 Th e w es t s lo pe a t t he T ia ns he ng qi ao H yd ro po w er St at io n bu ild in g w ea kl y w ea th er ed sa nd -s ha le m os ai c po or pr es pl itt in g sl op e st ru ct ur al c on tr ol 16 .1 35 6. 5 0. 25 0. 02 7 5 Th e w at er -in ta ke sl op e of th e D on g Fe ng H yd ro po w er S ta tio n w ea kl y w ea th er ed li m es to ne in fil l w ea k in te rc al at ed la ye r m os ai c po or pr es pl itt in g sl op e st ru ct ur al c on tr ol 60 45 9 0. 25 0. 02 5 6 Th e cr ee p de fo rm at io n bo dy a t t he m id dl e da m si te of th e M ia o Jia ba H yd ro po w er S ta tio n w ea kl y w ea th er ed m et a tu ff an d tu ffa ce ou s s la te ca ta cl as tic po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 11 8 45 7 0. 00 0. 02 6 7 Th e sl op e at th e Sh i M ol in g re se rv oi r a re a w ea kl y w ea th er ed sa nd st on e, sh al e m os ai c sa tis fa ct or y na tu ra l s lo pe st ru ct ur al c on tr ol 11 5 30 6. 5 0. 00 0. 02 7 8 Th e w at er -in ta ke sl op e of th e D a Ch ao sh an H yd ro po w er S ta tio n st ro ng ly w ea th er in g ba sa lt in fil l r hy ol ite a nd py ro cl as tic ro ck ca ta cl as tic po or na tu ra l s lo pe st ru ct ur al c on tr ol 71 25 25 0. 00 0. 02 8 9 Th e ab ut m en t s lo pe o n th e le ft b an k of th e La X iw a H yd ro po w er S ta tio n st ro ng ly - w ea kl y w ea th er in g gr an ite ca ta cl as tic po or na tu ra l s lo pe st ru ct ur al c on tr ol 80 25 32 0. 00 0. 02 5 10 Th e sl op e at th e N an Y i H yd ro po w er S ta tio n bu ild in g w ea kl y w ea th er in g po rp hy rit ic g ra ni te bl oc ky sa tis fa ct or y pr es pl itt in g sl op e st ru ct ur al c on tr ol 10 5 35 25 0. 25 0. 02 5 11 Th e sl op e at th e W o H us ha n H yd ro po w er S ta tio n sp ill w ay th ic k lim es to ne in fil l m ud st on e m os ai c po or pr es pl itt in g sl op e st ru ct ur al c on tr ol 43 45 9 0. 25 0. 02 6 12 Th e sl op e at th e re se rv oi r a re a of S an B an xi H yd ro po w er S ta tio n st ro ng ly w ea th er ed tu ffa ce ou s s ilt st on e m os ai c po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 50 45 7 0. 00 0. 02 5 13 Th e sl op e at th e in le t o f t he d iv er si on tu nn el a t Ta i P in gy i H yd ro po w er S ta tio n w ea kl y w ea th er ed g ra ni te bl oc ky po or co nv en tio na l bl as tin g st ru ct ur al c on tr ol 11 0 40 25 0. 50 0. 02 5 14 Th e sl op e at th e in le t o f t he d iv er si on tu nn el a t th e Li ji ax ia H yd ro po w er S ta tio n st ro ng ly w ea th er ed m ig m at ite in fil l s ch is t ca ta cl as tic po or sm oo th b la st in g st ru ct ur al c on tr ol 63 35 21 0. 50 0. 02 7 15 Th e sl op e at Ta i P in gy i H yd ro po w er S ta tio n ta ilr ac e st ro ng ly w ea th er ed g ra ni te m os ai c po or co nv en tio na l bl as tin g st ru ct ur al c on tr ol 11 0 35 25 0. 75 0. 02 5 16 Th e sl op e at Ta i P in gy i H yd ro po w er S ta tio n in ta ke w ea kl y w ea th er ed g ra ni te m os ai c sa tis fa ct or y co nv en tio na l bl as tin g st ru ct ur al c on tr ol 11 0 40 25 0. 75 0. 02 5 17 Th e sl op e at th e ex it of 3 # C av e of M an W an H yd ro po w er S ta tio n co m pl et el y st ro ng ly w ea th er ed rh yo lit e ca ta cl as tic sa tis fa ct or y co nv en tio na l bl as tin g no n- st ru ct ur al co nt ro l 90 30 25 0. 75 0. 02 6 18 Th e sl op e at th e Er ta n H yd ro po w er S ta tio n 2# ta ilr ac e w ea kl y w ea th er ed b as al t bl oc ky po or co nv en tio na l bl as tin g st ru ct ur al c on tr ol 12 0 45 16 0. 75 0. 02 8 G eologia C roatica Xiu-Zhen et al.: Modification of slope stability probability classification and its application to rock slopes in hydropower engineering regions 79 N um be r Sl op e na m e Li th ol og y Ro ck m as s st ru ct ur e di sc on tin ui ty co nd iti on Ex ca va tio n m et ho d Sl op e ty pe σ c (M Pa ) G SI m i D Γ （ M N /m 3 ） 19 Th e sl op e at th e Er ta n H yd ro po w er S ta tio n sp ill w ay in le t w ea kl y w ea th er ed b as al t bl oc ky sa tis fa ct or y co nv en tio na l bl as tin g st ru ct ur al c on tr ol 15 0 45 16 0. 75 0. 02 8 20 H e Jia la nd sl id e at th e M ia o Jia ba H yd ro po w er S ta tio n w ea kl y w ea th er ed m et a tu ff ca ta cl as tic po or na tu ra l s lo pe st ru ct ur al c on tr ol 11 8 30 8 0. 00 0. 02 4 21 D a H ua ng ya sl op e at th e W u jia ng du H yd ro po w er S ta tio n st ro ng ly w ea th er ed th ic k lim es to ne in fil l m ud st on e bl oc ky po or na tu ra l s lo pe st ru ct ur al c on tr ol 75 40 9 0. 00 0. 02 5 22 G u Sh iq un la nd sl id e at th e G on g Bo xi a H yd ro po w er S ta tio n w ea kl y w ea th er ed g ne is s lo os e po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 72 .9 25 28 0. 00 0. 02 7 23 Th e sl op e on th e m ou nt ai n be hi nd th e Ti an sh en gq ia o I H yd ro po w er S ta tio n bu ild in g sl ig ht ly -n ew ly m ed iu m th ic k m ud st on e in fil l s an ds to ne m os ai c sa tis fa ct or y un co nt ro lle d bl as tin g st ru ct ur al c on tr ol 36 50 11 0. 50 0. 02 5 24 Th e sl op e at th e Ti an sh en gq ia o I H yd ro po w er S ta tio n sp ill w ay sl ig ht ly w ea th er ed th ic k lim es to ne bl oc ky sa tis fa ct or y un co nt ro lle d bl as tin g st ru ct ur al c on tr ol 67 .5 55 10 0. 50 0. 02 5 25 Sh i P in gt ai la nd sl id e at th e Xi ao X ia sh i H yd ro po w er S ta tio n w ea kl y w ea th er ed m ig m at ite a nd m et as an ds to ne lo os e po or na tu ra l s lo pe st ru ct ur al c on tr ol 50 25 25 0. 00 0. 02 5 26 6# la nd sl id e at th e Ji Sh ix ia H yd ro po w er S ta tio n st ro ng ly w ea th er ed sa nd st on e an d co ng lo m er at e ca ta cl as tic po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 68 .4 35 17 0. 00 0. 02 6 27 Th e sl op e at th e M an W an H yd ro po w er S ta tio n st on e pi t w ea kl y- sl ig ht ly w ea th er ed rh yo lit e bl oc ky go od pr es pl itt in g sl op e st ru ct ur al c on tr ol 85 50 25 0. 25 0. 02 6 28 1# la nd sl id e at th e Ji Sh ix ia H yd ro po w er S ta tio n st ro ng ly w ea th er ed sa nd st on e an d co ng lo m er at e lo os e po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 67 20 17 0. 00 0. 02 6 29 Th e sl op e at th e Ti an sh en gq ia o II H yd ro po w er S ta tio n su rg e ta nk st ro ng ly w ea th er ed sa nd -s ha le a nd c ru sh ed st on e m os ai c go od pr es pl itt in g sl op e st ru ct ur al c on tr ol 32 .2 45 11 0. 25 0. 02 7 30 Th e st ee p cl iff sl op e on th e w es t s lo pe o f t he T ia ns he ng qi ao II H yd ro po w er S ta tio n st ro ng ly w ea th er ed sa nd st on e an d sh al e bl oc ky go od co nv en tio na l bl as tin g st ru ct ur al c on tr ol 32 .2 45 11 0. 75 0. 02 7 31 Th e w ed ge sl op e on th e rig ht b an k of th e M ia o Jia ba H yd ro po w er S ta tio n w ea kl y w ea th er ed m et as an ds to ne a nd tu ff bl oc ky po or na tu ra l s lo pe st ru ct ur al c on tr ol 11 8. 5 45 13 0. 00 0. 02 5 32 H ua ng L as hi la nd sl id e at th e Th re e G or ge s w ea kl y- sl ig ht ly w ea th er ed li m es to ne a nd c la ys to ne m os ai c po or na tu ra l s lo pe no n- st ru ct ur al co nt ro l 23 .6 9 45 8 0. 00 0. 02 6 33 Th e sl op e in fr on t o f t he P u Bu go u H yd ro po w er S ta tio n da m w ea kl y w ea th er ed b as al t bl oc ky sa tis fa ct or y na tu ra l s lo pe st ru ct ur al c on tr ol 90 50 25 0. 00 0. 02 8 34 Th e sl op e at th e Ti an sh en gq ia o II H yd ro po w er S ta tio n So ut h fa ct or y bu ild in g st ro ng ly w ea th er ed sa nd -s ha le bl oc ky po or pr es pl itt in g sl op e st ru ct ur al c on tr ol 11 .5 45 12 0. 25 0. 02 8 Ta bl e A1 . C on tin ue d. G eo lo gi a C ro at ic a Geologia Croatica 72 / Special Issue80 Ta bl e A2 . T he a na ly si s r es ul ts o f t he m od ifi ed S SP C sy st em fo r t he 3 4 ro ck sl op es . N um be r Sl op e he ig ht H (m ) Sl op e an gl e β s (° ) c′ (M Pａ ) φ ′( °) H m ax (m ) H m ax /H φ ′/ β s St ab ili ty p ro ba bi lit y SP Ac tu al st ab ili ty 1＊ 20 0 35 0. 71 17 .2 0 13 3. 57 0. 67 0. 49 5% un st ab le 2＊ 25 5 32 1. 86 30 .5 2 23 6. 58 0. 93 0. 95 35 % un st ab le 3＊ 21 0 32 1. 86 30 .5 2 25 2. 35 1. 20 0. 95 57 % un st ab le 4 12 4 42 0. 42 18 .7 1 38 .1 8 0. 31 0. 45 3% pa rt ia lly u ns ta bl e 5 15 6 49 2. 40 25 .9 1 30 8. 22 1. 98 0. 53 85 % pa rt ia lly u ns ta bl e 6＊ 35 0 39 4. 80 26 .0 0 48 4. 95 1. 39 0. 67 92 % un st ab le 7 22 0 50 4. 61 25 .5 0 21 6. 28 0. 98 0. 51 39 % un st ab le 8 85 40 3. 12 30 .8 0 37 1. 01 4. 36 0. 77 97 % pa rt ia lly u ns ta bl e 9 25 0 50 3. 98 33 .6 2 61 8. 79 2. 48 0. 67 97 % un st ab le 10 15 0 73 6. 84 37 .9 7 92 4. 12 6. 16 0. 52 98 % st ab le 11 13 2 42 1. 72 25 .9 0 21 2. 40 1. 61 0. 62 96 % pa rt ia lly u ns ta bl e 12 ＊ 38 0 40 2. 05 26 .1 1 21 3. 71 0. 56 0. 65 4% un st ab le 13 12 0 70 6. 99 37 .4 7 11 87 .2 1 9. 89 0. 54 99 % pa rt ia lly u ns ta bl e 14 13 5 43 2. 39 25 .8 5 42 7. 41 3. 17 0. 60 97 % pa rt ia lly u ns ta bl e 15 11 0 35 4. 93 28 .9 6 96 8. 12 8. 80 0. 83 97 % pa rt ia lly u ns ta bl e 16 82 70 4. 93 28 .9 6 11 87 .2 1 14 .4 8 0. 41 10 0% st ab le 17 ＊ 17 0 45 2. 52 20 .1 7 62 1. 08 3. 65 0. 45 98 % un st ab le 18 11 0 52 4. 22 23 .4 2 91 6. 26 8. 33 0. 45 99 % pa rt ia lly u ns ta bl e 19 90 64 .5 4. 22 23 .4 2 11 45 .3 2 12 .7 3 0. 36 10 0% pa rt ia lly u ns ta bl e 20 88 0 43 3. 83 22 .8 4 28 9. 97 0. 33 0. 53 3% un st ab le 21 30 0 60 3. 06 26 .7 0 30 6. 73 1. 02 0. 45 45 % pa rt ia lly u ns ta bl e 22 ＊ 44 5 23 3. 33 31 .8 0 44 9. 00 1. 01 1. 38 70 % st ab le 23 14 0 45 1. 49 26 .4 4 27 3. 37 1. 95 0. 59 98 % st ab le 24 13 0 63 2. 99 27 .5 7 59 3. 92 4. 57 0. 44 96 % st ab le 25 11 9 35 2. 20 30 .8 0 29 2. 63 2. 46 0. 88 94 % st ab le 26 ＊ 33 0 45 2. 62 27 .4 4 38 7. 15 1. 17 0. 61 85 % st ab le 27 19 3 73 5. 97 39 .7 0 13 26 .5 1 6. 87 0. 54 98 % st ab le 28 21 8 31 2. 56 27 .4 4 19 8. 37 0. 91 0. 89 40 % un st ab le 29 12 0 43 1. 38 27 .5 3 18 0. 78 1. 51 0. 64 94 % st ab le 30 11 0 63 .4 0. 98 20 .6 7 18 0. 78 1. 64 0. 33 96 % pa rt ia lly u ns ta bl e 31 25 0 42 5. 92 31 .2 6 83 3. 75 3. 34 0. 74 97 % st ab le 32 ＊ 80 0 30 1. 01 27 .2 0 10 6. 98 0. 13 0. 91 2% un st ab le 33 30 0 37 5. 91 38 .4 5 13 04 .2 2 4. 35 1. 04 80 % st ab le 34 70 51 0. 51 28 .2 5 67 .2 0 0. 96 0. 55 35 % pa rt ia lly u ns ta bl e N ot e: “* ” i s t he sl op e of n on -s tr uc tu ra l c on tr ol fa ilu re .